Scans, stamps and photos

What this covers¶

Four tools for the papers and pictures that travel with a project: OCR reads a scan, Stamp a Signature puts your signature on a PDF, Background Remover cuts a subject out of a photo, and EXIF Extractor turns geotagged photos into points on a map. This page also says how the PDF tools handle Arabic.

All of them run on the Haiyz processing engine, not in your browser: the file is uploaded, processed, and the result is yours to download or keep.

OCR — reading a scan¶

OCR reads the text of a scanned page in English, Arabic, or both together.

What you can send

• An image: PNG, JPEG, WebP, BMP or TIFF.
• A PDF of scanned pages, up to 30 pages. Each page is read in turn.

Options

Option	Choices	Use it when
Language	English, Arabic, English + Arabic	Choose both only when the page really mixes them; a single language reads better
Page layout	A page of text · One block · Scattered text	Scattered text suits a map sheet, a form or a label
Result	Text · Text and a searchable PDF	The searchable PDF is your scan with its text underneath, so it can be searched and copied
Turn the page	On or off	On by default: a page scanned sideways or upside down is turned before it is read

What comes back

• The text, with pages separated.
• Confidence, from 0 to 100, for each page and overall. It is how sure the reading is, not a promise that every word is right.
• A warning when the confidence is low. The text is then likely wrong in places; a sharper, larger or better-lit scan reads better.
• How many pages were turned upright.
• The searchable PDF, when you asked for it.

Getting a good reading

• Scan at 300 dpi where you can. A phone photo works when the page is flat, sharp and evenly lit.
• A small or faint page is enlarged and its contrast stretched before it is read. That helps, but it cannot recover detail the photo never had.
• Handwriting, stamps over text and decorative Arabic scripts read poorly.

A PDF that already has selectable text does not need OCR. PDF to Word tells you when a PDF is a scan; this tool is where to take it.

Stamp a Signature¶

This tool places an image of your signature on the pages of a PDF, at the positions you click. You can draw the signature, upload a PNG, or reuse one you saved.

It is a stamp, not a certified digital signature. It carries no certificate, no PDF viewer will verify it, and it does not detect later changes to the file. Use it where a visible signature is what is asked for.

• Name under the stamp writes one line beneath each stamp: your name and, if you choose, today's date. Arabic names are written as real text.
• If the PDF was digitally signed, stamping changes the file, so that signature no longer verifies in the stamped copy. The tool makes the stamp and tells you; your original file is not changed.
• The result is named with -stamped.

Background Remover¶

Removes what is behind the subject of a photo.

Option	What it does
Result	The subject, or the mask alone — white where the subject is — for use in an editor
Behind the subject	Nothing (a transparent PNG), white, black, or a colour you give as a hex code such as #1E88E5
Clean the edge	Removes specks and smooths a ragged outline
Keep soft edges	For hair, fur and other fine edges. Slower

It works best on one clear subject against a background that differs from it. The first run after the service starts takes a few seconds longer.

EXIF Extractor — photos as points¶

Upload up to 10 JPEG photos. For each one the tool reads where and when it was taken and returns:

• its position, as longitude and latitude;
• its altitude in metres — negative below sea level;
• the direction the camera faced, in degrees clockwise from north, and whether that north is true or magnetic;
• the camera;
• a clean copy of the photo with all metadata removed.

The geotagged photos come back together as a point layer (GeoJSON) carrying those values, ready to open in the Workspace: photos in, a layer on the map. A photo without a position is listed and left out of the layer.

Arabic in PDFs¶

The PDF tools write Arabic as real text — shaped, right to left, and selectable:

• PDF Watermark: an Arabic watermark, or one that mixes Arabic and Latin such as سري CONFIDENTIAL, is drawn as text at the opacity and angle you choose.
• Fill PDF Form: an Arabic value typed into a form field shows correctly, whether or not you flatten the form.
• Stamp a Signature: the name under the stamp.

From the API¶

Each option is a field of the tool's request; the tool reference lists them. OCR answers JSON with text, confidence, pageCount and, when asked, pdfKey — the key of the searchable PDF to download. A request that sends only a file is answered as before.